Action-Aware Generative Sequence Modeling for Short Video Recommendation

2026-04-28

AI Summary

This work addresses the limitation of existing short-video recommendation models, which treat videos as monolithic units and thus fail to capture users’ fine-grained preferences toward internal segments. To overcome this, the authors propose an action-aware generative sequential recommendation framework that uniquely integrates fine-grained user action temporal patterns with contextual features. The approach employs a context-aware attention module (CAM), a hierarchical sequential encoder (HSE), and an action-sequence autoregressive generator (AAG) to enable unified modeling of user behavior dynamics. Online A/B experiments demonstrate significant improvements in user engagement: watch time increases by 0.34%, interaction rate rises by 8.1%, and seven-day retention improves by 0.162%. The system has been fully deployed and currently serves over 400 million users.

Abstract

With the rapid development of the Internet, users have increasingly higher expectations for the recommendation accuracy of online content consumption platforms. However, short videos often contain diverse segments, and users may not hold the same attitude toward all of them. Traditional binary-classification recommendation models, which treat a video as a single holistic entity, face limitations in accurately capturing such nuanced preferences. Considering that user consumption is a temporal process, this paper demonstrates that the timing of user actions can represent diverse intentions through statistical analysis and examination of action patterns. Based on this insight, we propose a novel modeling paradigm: Action-Aware Generative Sequence Network (A2Gen), which refines user actions along the temporal dimension and chains them into sequences for unified processing and prediction. First, we introduce the Context-aware Attention Module (CAM) to model action sequences enriched with item-specific contextual features. Building upon this, we develop the Hierarchical Sequence Encoder (HSE) to learn temporal action patterns from users' historical actions. Finally, through leveraging CAM, we design a module for action sequence generation: the Action-seq Autoregressive Generator (AAG). Extensive offline experiments on the Kuaishou's dataset and the Tmall public dataset demonstrate the superiority of our proposed model. Furthermore, through large-scale online A/B testing deployed on Kuaishou's platform, our model achieves significant improvements over baseline methods in multi-task prediction by leveraging sequential information. Specifically, it yields increases of 0.34% in user watch time, 8.1% in interaction rate, and 0.162% in overall user retention (LifeTime-7), leading to successful deployment across all traffic, serving over 400 million users every day.

Problem

• short video recommendation
• user preference modeling
• temporal action patterns
• fine-grained interaction
• sequence modeling

Innovation

• Action-Aware Modeling
• Generative Sequence Network
• Temporal Action Patterns
• Context-aware Attention
• Short Video Recommendation